INTERCORR2010_399

Copyright 2010, ABRACO Trabalho apresentado durante o INTERCORR 2010, em Fortaleza/CE no mês de maio de 2010. As informações e opiniões contidas neste trabalho são de exclusiva responsabilidade do(s) autor(es).

  

EVALUATION OF ANTICORROSIVE COATINGS FOR TANKER WALLS TO TRANSPORT OIL IN A HIGH SALINITY ENVIRONMENT, IN THE PRESENCE OF

TENSION, TEMPERATURE AND CO2

Neusvaldo Lira de Almeida1; Adriano Garcia Bernal1; Victor Solymossy2; Flávio

Augusto S. Serra2; Joaquim Pereira Quintela2

ABSTRACT The offshore heavy oil project uses Floating Production, Storage and Offloading

(FPSO) structural tanks to act as oil separators (washing tanks) a vital requirement.

These tanks operate with a continuous layer of produced water at high operational

temperatures and high residence time, creating a critical corrosive environment.

This condition is far beyond typical conditions for tankers and makes the

development of a special test necessary to qualify paint systems for that condition.

Traditionally the qualification of paint systems in Brazil offshore units is based on

NACE TM-0104 and ISO 20340 standards. But these standards have a test protocol

based on physical/mechanical tests and corrosion tests. However, for this specific

case, it did not suffice. Thus, a specific test was developed for washing tanks, which

can evaluate simultaneously both corrosion and mechanical properties. This paper

presents the development of a new test to simulate the behavior of paint systems

submitted to cyclic tension in coatings in very corrosive environments (high salinity,

CO2, and high temperature).

Keywords: tankers, coatings, fatigue, heavy oil, salinity.

1 Pesquisador do Laboratório de Corrosão e Proteção do Instituto de Pesquisas Tecnológicas do

Estado de São Paulo; 2 Pesquisador da Petrobrás - Cenpes

INTERCORR2010_399

- 2 -

INTRODUCTION FPSO structural cargo tanks are not rigid structures. Shell plates are somewhat

flexible and subjected to cyclical loads due to wave and cargo loading and

offloading. This situation associated to very corrosive environment (high

temperature, presence of carbon dioxide and high salinity of the produced water) is

a challenge when a paint system to protect the carbon steel against corrosion must

be chosen. Corrosion tests only were not considered enough, since the coating is

subjected to high stress due the tank shell movements (some projects consider that

the material must withstand tensile stress up to 80% of yield point). Consequently,

the anticorrosion paint system must have sufficient flexibility and fatigue resistance

in addition to good corrosion resistance.

Usually the test protocols like NACE TM -104, NACE TM-0404 and ISO 20340 have

distinct tests for evaluating corrosion and mechanical/physical properties but not

one evaluate at the same time both corrosion and mechanical properties. It was

found that, for washing tanks, It could be more representative to test simulating both

an aggressive environment and tensile tests the paint could be subjected during

operation. With this goal, a test procedure based on cyclic stress condition plus a

corrosive environment was developed.

TEST PROTOCOL CONCEPT The developed test protocol concept was based on three-point flexural test applying

a cycling load onto a painted carbon steel plate. The cyclic flexural load simulates

the tankers wall deflections and subjects the paint system to the tension levels

withstand by the tank walls.

Tension level applied must be representative of the worst conditions, so we

selected the tension level which can stress the base material at 80% of material´s

yield strength, and the number of cycles was select to 1,000,000 (considered a

lifetime for a typical FPSO).

The corrosive environment means high salinity and high temperature (60ºC). Initially

H2S saturated solution was considered, but the costs and difficulties of managing

INTERCORR2010_399

- 3 -

H2S could made the test protocol not practical and, more important, it could greatly

increase the cost of the test protocol (the aim was, since the beginning, a low cost

test protocol available to the paint manufacturers). So we changed to CO2 saturated

solution an aggressive reagent also found in oil fields.

This work presents new test to simulate the tensile stress in coating and to evaluate

the fatigue resistance associated to very corrosive environments as: high salinity,

CO2, and high temperature in coatings applied on the walls of the tankers. The

results can supply more appropriate information.

The test, nominated “Fatigue Resistance Test in Coatings”, was carried out in

synthetic seawater saturated with carbon dioxide at 60°C. Four different coatings

were studied: (1) epoxy with glass flakes; (2) thermal spray aluminum (TSA), sealed

with epoxy mist coat; (3) thermal spray aluminum (TSA), sealed with epoxy mist

coat plus an additional layer of epoxy novolac with glass flakes and (4) thermal

spray zinc-aluminum (TSAZ), sealed with epoxy mist coat plus an additional layer of

epoxy with glass flakes, all of them applied on carbon steel panels.

METHODOLOGY USED TO CALCULATE THE TENSION The corrected applied tensile level was a great concern. The first step was to made

tensile tests to discover the yield point of uncovered carbon steel. Another

challenge was the methodology to apply the tension. The methodology used to

apply the tension in the sample was based on a three-point device in agreement

with ASTM G 39, shown in Figure 1. There is a relationship between the deflection

and the applied tension that is given by Equation (1). In this arrangement, the

maximum tension occurs exactly at the middle point of the distance between the

supporting points and it decreases linearly to nil at the edge.

INTERCORR2010_399

- 4 -

FIGURE 1 - Schematic representation of the three-point device used to define the relationship between deflection and tension.

σ (1)

Where:

σ = maximum tensile stress;

E = modulus of elasticity of the test material;

t = thickness of the sample;

y = maximum deflection;

H = distance between supporting points of the sample.

Equation (1) is valid for small deflections, typically for y/H smaller than 0.1.

According to this equation, once the yield point of the material as well as parameter

E; t and H are known, it is possible to apply different deflection levels corresponding

to different percentages of yield point of the material.

The sample of carbon steel used as a substrate was tested to determine the yield

point of the material. Then the deflection corresponding to 80 % of yield point was

calculated, using Equation (1).

To assess the validity of these assumptions, a strain gauge was coupled on the

metallic plate made, with the same material and with the same dimensions as the

sample to be tested, and the sample subject a deflection corresponding to 80% of

the yield point of the material. The monitored areas of the sample are illustrated in

Figure 2 and they correspond to 80 %, 68 %, 56 % and 35% of the yield point of the

INTERCORR2010_399

- 5 -

material. The relationship between the deflection and the tension is shown in the

Figure 3 and confirms the linear relationship assumption.

EQUIPMENT The equipment was developed by Laboratory of Corrosion and Protection of

Institute for Technological Research of Sao Paulo State. Figure 2 shows the

equipment. Some components of the equipment are presented below and in Figure

3:

• Tank 50 cm width, 45 cm length and 30 cm depth;

• The up-and-down mechanism to apply cyclical deflection;

• The electric resistance to warm the solution. The resistance was made of

resistant material against corrosive environment;

• Rocker arm with mandrel of stainless steel with adjustable cable to apply

cyclical tension. The mandrel width is similar to the sample width, for good

tension uniformity;

• The frequency control system;

• The system for gas injection, with a C-ring of stainless steel. The C-ring

contains some holes to improve the gas distribution inside the tank;

• Stainless steel support for holding the sample;

• The cycle counter;

• The thermocouple;

• The temperature recorder;

• The graduated meter to adjust the deflection in the sample.

INTERCORR2010_399

- 6 -

FIGURE 2 - Equipment developed for fatigue resistance test in coatings.

FIGURE 3 – Parts of the equipment.

INTERCORR2010_399

- 7 -

SAMPLES Carbon steel panels SAE 1020, 30 cm length; 10 cm width and 2 mm thick were

blasted with bauxite up to the visual standard Sa 2 ½ of ISO 8501-1. Four different

coatings were studied; all of them applied on carbon steel panels.

• Coating A – Thermal Spray Aluminum – TSA, 200 µm thickness sealed with

epoxy mist coat;

• Coating B – Epoxy novolac with glass flakes, 900 µm DFT;

• Coating C – Thermal Spray Aluminum – TSA, 200 µm thickness plus epoxy

novolac with glass flakes, 900 µm DFT;

• Coating D – Thermal Spray Aluminum-Zinc – TSAZ, 200 µm thickness plus

epoxy novolac with glass flakes, 900 µm DFT.

CHARACTERIZATION TESTS Initially, the samples were submitted to Holiday Detector test and to Pull-Off test.

Holiday Detector Test was done in agreement with ASTM G 62(07) - Standard Test

Methods for Holiday Detection in Pipeline Coatings. In this test method, an

electrode was positioned on the coating and another terminal was connected at a

substrate to make electric contact. Then 5,000 V was applied between the electrode

and the substrate. A characteristic sound indicates that there are pores or cracking

in the coating.

Pull-Off test was done in agreement with ASTM D 4541(09) using an epoxy based

adhesive system with a bonding strength to 3,000 PSI. This test method uses a

class of apparatus known as portable pull-off adhesion testers. They are capable of

applying a concentric load and counter load to a single surface so that coatings can

be tested even though only one side is accessible.

TEST CONDITIONS 50 liters of synthetic seawater were used. The water was prepared in agreement

with ASTM D1141 (see chemical composition in Table 1). It was saturated with

INTERCORR2010_399

- 8 -

carbon dioxide during the test. The agitation was done by carbon dioxide bubbles

and by oscillatory movement of the sample.

It is known that the salinity of the petroleum-water is higher than the synthetic sea

water. Even so, it was used because it is a standardized water and it would enable

repeating the test whenever desired.

The test solution temperature was heated to 60 ºC using an electric immersion

heater and measured with a thermo-couple installed close and parallel to sample.

The test duration was 1,000,000 cycles and the oscillation frequency was 2 Hz. The

pH of the solution was monitored. Initial pH was 7.5 e after CO2 injection stabilized

in 5.1.

TABLE 1 Composition of synthetic sea water (ASTM D 1141)

Reagents Concentration (grams per liter)

NaCl 24.53

MgCl2 5.20

Na2SO4 4.09

CaCl2 1.16

KCl 0.695

NaHCO3 0.201

KBr 0.101

H3BO3 0.027

SrCl2 0.025

NaF 0.003

APPLIED TENSION

The maximum deflection applied on the sample corresponded to 80% of the yield

point of the material. The deflection was calculated according to Equation (1) and

the adjustment was made through a scale of 0.01 mm resolution.

INTERCORR2010_399

- 9 -

Supposedly the coating would not be able to withstand the deflection corresponding

to 80% of the yield point of the material, another four areas of the sample were

monitored with strain gauges. Thus, the highest deflection withstood by the coatings

could be determined. Monitored areas of the sample are illustrated in Figure 4 and

they correspond to 80%, 68%, 56% and 35% of the yield point of the material. The

relationship between the deflection and the tension is shown in Figure 5.

Region A Region B Region C Region D

FIGURE 4 - Schematic representation of the monitored positions of the samples.

FIGURE 5 - Relationship between the deflection and the percentage of the yield point of the substrate material.

0

0,2

0,4

0,6

0,8

1

1,2

0 5 10 15

%  of Y

P MPa

)

Deflection in mm

A

B

C

D

80 % of yp

68 % of yp

56 % of yp

35 % of yp

TESAfter

with

not p

novo

TSZ

insp

show

are p

Coa

•

•

•

FI

ST RESULTr the test,

epoxy pre

possible to

olac with g

A sealed p

ections we

w the coat

presented

ting A – T

Pull-Off te

Holiday D

and no cr

Visual ex

from alum

was not

(see Figu

IGURE 6 -

TS a visual in

esented w

o evaluate

glass flake

plus epoxy

ere done u

tings after

as follows

Thermal sp

est before

Detector te

racking;

xamination

minum wer

possible to

ure 6).

Coating Atest. W

spection w

hite corros

e other def

es, TSA se

y novolac w

using Holli

Pull-Off te

.

pray alum

the fatigue

est before t

after the

re observe

o evaluate

A - TSA seWhite corro

with 40x ma

sion produ

fects in the

ealed plus

with glass f

iday Detec

est and af

inum – TS

e resistanc

the fatigue

fatigue re

ed due to

e the crack

ealed withosion prod

agnificatio

ucts becau

e coating.

s epoxy no

flakes, no d

ctor and P

fter fatigue

SA Sealed

ce test: 7.0

resistance

sistance te

low thickn

king and p

h epoxy aftducts were

IN

n was don

se of its th

In the oth

ovolac with

damage wa

Pull-off test

e resistanc

d With Epo

0 MPa (only

e test, with

est: white

ess of sea

pore forma

fter the fate observe

NTERCORR

ne. The TS

hinness; s

her coating

h glass fla

as observe

ts. Figures

ce test. Th

oxy

y TSA);

h 5,000 V:

corrosion

aling layer

ation in the

tigue resisd.

R2010_399

- 10 -

SA sealed

so, it was

gs: epoxy

akes and

ed. Other

s 6 to 12

e results

no pores

products

. Thus, it

e coating

stance

INTERCORR2010_399

- 11 -

Coating B – Epoxy Novolac With Glass Flakes

• Pull-Off test before the fatigue resistance test: 5.4 MPa - cohesive failure;

• Pull-Off test after the fatigue resistance test: 5.4 MPa – cohesive failure

(Figure 7);

• Holiday Detector test before the fatigue resistance test, with 5,000 V: no pores

and no cracking;

• Holiday Detector after the fatigue resistance test, with 5,000 V: no pores and

no cracking;

• Visual examination after the fatigue resistance test: there was no cracking or

any other defects in the coating (Figure 8).

FIGURE 7 – Coating B – Epoxy novolac with glass flakes. Pull-

Off test, after fatigue resistance test – cohesive failure.

Coa

Flak

•

•

•

•

•

FIGURE re

ting C – T

kes

Pull-Off te

Pull-Off te

novolac w

Holiday D

and no cr

Holiday D

and no cr

Visual ex

other def

8: Coatingesistance

Thermal S

est before

est after th

with glass f

Detector te

racking;

Detector te

racking;

xamination

fect in the c

g B – Epo test. The

Spray Alum

the fatigue

he fatigue r

flakes) (Fig

est before t

est after th

after fatig

coating (Fi

oxy novolare was no

minum (T

e resistanc

resistance

gure 9);

the fatigue

he fatigue r

ue resista

gure 10).

ac with glao cracking

TSA) Plus

ce test: 4.4

test: 4.4 M

resistance

resistance

nce test: th

IN

ass flakesin the coa

Epoxy No

4 MPa – co

MPa – cohe

e test, with

test, with

here was n

NTERCORR

after fatigating.

ovolac Wi

ohesive fail

esive failur

h 5,000 V:

5,000 V:

no crackin

R2010_399

- 12 -

gue

th Glass

ure;

re (epoxy

no pores

no pores

g and no

FI

Coa

Glas

•

•

FIGUREPull

IGURE 10fatigu

ting D – T

ss Flakes

Pull-Off te

Pull-Off t

(Figure 1

E 9: Coati-Off test, a

: Coating ue resistan

Thermal Sp

The result

est, before

test, after

1);

ng C – TSafter fatig

C – TSA pnce test. T

pray Zinc-

ts are pres

e the fatigu

the fatigu

SA plus ep

ue resista

plus epoxyThere was

-Aluminum

ented as fo

e resistanc

ue resistan

poxy novoance test –

y novolacno cracki

m (TSAZ) P

ollows.

ce test: 4.4

nce test:

IN

lac with g– cohesive

c with glasing in the

Plus Epox

4 MPa – co

4.4 MPa

NTERCORR

glass flakee failure.

ss flakes, acoating.

xy Novola

ohesive fai

– adhesiv

R2010_399

- 13 -

es.

after

c With

ilure;

ve failure

•

•

•

FIG

Holiday D

pores and

Holiday D

and no cr

Visual ex

other def

FIGURflakes. P

GURE 12: Cfatigu

Detector t

d no crack

Detector te

racking;

xamination

fect in the c

RE 11 – CoPull Off tes

Coating Due resista

est, before

king;

est after th

after fatig

coating (Fi

oating D -st, after fa

– TSZA pance test. T

e the fatig

he fatigue r

ue resista

gure 12).

- TSZA pluatigue resi

plus epoxyThere was

gue resista

resistance

nce test: th

us epoxy nistance tes

y novolac s no crack

IN

ance test,

test, with

here was n

novolac wst – adhes

with glassking in the

NTERCORR

with 5,00

5,000 V:

no crackin

with glass sive failur

s flakes, ae coating.

R2010_399

- 14 -

00 V: no

no pores

g and no

re.

after the

INTERCORR2010_399

- 15 -

CONCLUSIONS The test and the methodology were considered appropriate to evaluate paint

systems subjected to cyclical tension and harsh corrosive environment (high

salinity, high temperature and carbon dioxide environment). In this test method,

coatings: B – epoxy novolac with glass flakes; C – Thermal Spray Aluminum plus

epoxy novolac with glass flakes and D – Thermal Spray Aluminum-Zinc plus epoxy

novolac with glass flakes presented good performances for cyclical deflection up to

80% of yield point of the substrate material. Coating A – Thermal Spray Aluminum

sealed with epoxy mist coat did not present a satisfactory performance. It was

observed white corrosion products from aluminum, probably due to low thickness of

sealing layer.

REFERENCES ASTM G 39 - Standard Practice for Preparation and Uses of Bent-Beam Stress-

Corrosion Test Specimens

ASTM G 62(07) – Standard Test Methods for Holiday Detection in Pipeline Coatings

ASTM D1141 – Standard Practice for the Preparation of Substitute Ocean Water ACKNOWLEDGMENT The authors acknowledge Petrobras Company by the financial sponsor.